Phylogeny and taxonomy of the North American clade of the ... · pirilliformis I. Barnes & M. J. Wingf. from Australia (Barnes et al 2003) and Africa (Roux et al 2004) and C. polychroma

Phylogeny and taxonomy of the North American clade of the Ceratocystisfimbriata complex

Jason A. Johnson1

Thomas C. Harrington2

C.J.B. EngelbrechtDepartment of Plant Pathology, Iowa State University,Ames, Iowa 50010

Abstract: Ceratocystis fimbriata is a widely distributed,plant pathogenic fungus that causes wilts and cankerson many woody hosts. Earlier phylogenetic analyses ofDNA sequences revealed three geographic cladeswithin the C. fimbriata complex that are centeredrespectively in North America, Latin America andAsia. This study looked for cryptic species within theNorth American clade. The internal transcribedspacer regions (ITS) of the rDNA were sequenced,and phylogenetic analysis indicated that most isolatesfrom the North American clade group into four host-associated lineages, referred to as the aspen, hickory,oak and cherry lineages, which were isolated primarilyfrom wounds or diseased trees of Populus, Carya,Quercus and Prunus, respectively. A single isolatecollected from P. serotina in Wisconsin had a uniqueITS sequence. Allozyme electromorphs also werehighly polymorphic within the North American clade,and the inferred phylogenies from these data werecongruent with the ITS-rDNA analyses. In pairingexperiments isolates from the aspen, hickory, oak andcherry lineages were interfertile only with otherisolates from their respective lineages. Inoculationexperiments with isolates of the four host-associatedgroupings showed strong host specialization byisolates from the aspen and hickory lineages onPopulus tremuloides and Carya illinoensis, respectively,but isolates from the oak and cherry lineages did notconsistently reveal host specialization. Morphologicalfeatures distinguish isolates in the North Americanclade from those of the Latin American clade(including C. fimbriata sensu stricto). Based on thephylogenetic evidence, interfertility, host specializa-tion and morphology, the oak and cherry lineages arerecognized as the earlier described C. variospora, thepoplar lineage as C. populicola sp. nov., and thehickory lineage as C. caryae sp. nov. A new speciesassociated with the bark beetle Scolytus quadrispinosus

on Carya is closely related to C. caryae and isdescribed as C. smalleyi.

Key words: Microascales, Scolytus quadrispinosus,speciation

INTRODUCTION

Species of Ceratocystis are largely insect-dispersedpathogens of woody plants, infecting their hoststhrough wounds. Ceratocystis fimbriata Ellis & Halstedis notable for its broad host range, with at least 31plant species from 14 families confirmed as hosts(CABI 2001). Hosts of C. fimbriata include Eucalyptusspp., Mangifera indica (mango), Theobroma cacao(cacao), Coffea arabica (coffee), Hevea brasiliensis(rubber tree), Platanus spp. (sycamore or plane tree),Prunus spp. (almond and other stone fruits) andPopulus spp. (aspen and other poplars). Nonwoodyhosts include Colocasia esculenta (taro) and Ipomoeabatatas (sweet potato), from which the speciesoriginally was described (Halsted 1890). The geo-graphic range and genetic diversity of C. fimbriata issimilarly impressive, although most of the diversity inthe species is found in the Americas (Baker et al 2003,Barnes et al 2001, CABI 2001, Harrington 2000,Steimel et al 2004). Ceratocystis albifundus, a closelyrelated species, is native to Africa (Roux et al 2000,Wingfield et al 1996).

Webster and Butler (1967) concluded that in-terfertility and lack of morphological differencesprecluded recognition of additional species withinC. fimbriata, but sequences of the internal transcribedspacer region (ITS) of the nuclear ribosomal DNAand other genetic analyses show that there are severalsubgroups or clades within C. fimbriata (CABI 2001,Harrington 2000). One of these major clades seems tobe centered in Latin America, where C. fimbriatainfects numerous native and nonnative hosts. Two ofthe members of the Latin American clade, the sweetpotato pathogen C. fimbriata sensu stricto and thesycamore pathogen C. platani (Walter) Engelbrecht& Harrington, are found in eastern North Americaand elsewhere (Baker et al 2003, Engelbrecht et al2004, Engelbrecht and Harrington 2005). Othermembers of the Latin American clade include thecacao pathogen C. cacaofunesta Engelbrecht &Harrington, a Xanthosoma pathogen in the Caribbe-an, and various Central and South American popula-tions (Baker et al 2003, Engelbrecht and Harrington

Accepted for publication 1 July 2005.1 Current address: Utah Division of Forestry Fire & State Lands, SaltLake City, Utah2 Corresponding author. E-mail: [email protected]

Mycologia, 97(5), 2005, pp. 1067–1092.# 2005 by The Mycological Society of America, Lawrence, KS 66044-8897

1067

TABLE I. Origin and GenBank accession numbers for ITS r-DNA sequences of Ceratocystis fimbriata isolates from the NorthAmerican clade

Lineagea

Isolate orspecimennumberb

GenBankAccession No.

Additional isolate (specimen)numbers/collectorc Source

CollectionLocation

Aspen C89 AY 907027 CBS 114725/Hinds Populus tremuloides SouthDakota

C685 AY 907028 CBS 11561, BPI 843723/Smalley

P. tremuloides Quebec

C947 ATCC 36291/ Gremmen Populus sp., from canker PolandC995 AY 907029 CBS 119.78/ Gremmen Populus sp., from canker PolandC1485 ATCC 24096/Hinds P. tremuloides Colorado

Hickory A C682 Smalley Carya cordiformis,Scolytus-infested

Wisconsin

C683 Smalley Carya ovata, fromScolytus

Wisconsin

C684 AY 907030 CBS 114724, BPI 843722/Smalley

C. cordiformis,Scolytus-infested

Wisconsin

C1410 AY 907031 Harrington C. cordiformis IowaC1411 Harrington C. cordiformis IowaC1828 AY 907032 Johnson C. cordiformis, with

ScolytusIowa

C1839 Johnson C. cordiformis,Scolytus-infested

Iowa

C1840 Johnson C. cordiformis, Scolytusbeetle gallery

Iowa

C1842 Johnson C. cordiformis IowaC1844 Johnson C. cordiformis, Scolytus

beetle galleryIowa

C1952d Johnson C. cordiformis, Scolytusbeetle gallery

Iowa

Hickory B C1412 AY 907033 BPI 843728/Harrington C. cordiformis, fromwound

Iowa

C1413 Harrington C. cordiformis IowaC1827 AY 907034 CBS 115168/Johnson Carya ovata, wound IowaC1829 AY 907035 CBS 114716, BPI 843735/

JohnsonC. cordiformis, wound Iowa

C1845 Johnson C. cordiformis, wound IowaC1971d Johnson Ostrya virginiana, wound Iowa

Oak C1009 CBS 773.73, ATCC 12861/Campbell

Quercus sp., from freshstump

Minnesota

C1483 AY 907036 ATCC 12866/ Campbell Quercus ellipsoidalis,from fresh stump

Minnesota

C1843 AY 907037 CBS 114715, BPI 843737/Johnson

Q. alba, wound Iowa

C1846 AY 907038 CBS 114714, BPI 843738/Johnson

Q. robur, bleedingcanker

Iowa

BPI 595631c AY 907039 Davidson Q. prinus, from innerbark

WestVirginia

Oak, Japan C1709 AY 907040 MCC-NIES 323/ Matsuya Betula platyphylla,from log

Japan

Cherry C578 AY 907041 Bostock Prunus dulcis, canker CaliforniaC686 Smalley P. dulcis CaliforniaC821 Rizzo P. dulcis, canker CaliforniaC855 Harrington P. dulcis, canker CaliforniaC856 Harrington P. dulcis, canker CaliforniaC857 Harrington P. dulcis, canker CaliforniaC1821 Harrington P. dulcis, canker California

1068 MYCOLOGIA

2005, Harrington 2000, Marin et al 2003, Thorpe et al2005). A second clade occurs on fig (Ficus carica) andtaro in Japan and the Pacific (Harrington 2000,Thorpe et al 2005), and the recently described C.pirilliformis I. Barnes & M. J. Wingf. from Australia(Barnes et al 2003) and Africa (Roux et al 2004) andC. polychroma M. van Wyk, M. J. Wingf. & E.C.Y. Liewfrom Indonesia (van Wyk et al 2004) are also in theAsian clade based on rDNA sequence analysis(Harrington unpublished, Thorpe et al 2005). Mem-bers of a third clade infect Populus spp., Carya spp.,Quercus spp., and Prunus spp. in North America(Harrington 2000). A morphologically similar species,C. variospora (R.W. Davidson) C. Moreau, was de-scribed from Quercus in the eastern USA (Davidson1944, Hunt 1956), but some have considered thisspecies and Rostrella coffaea to be synonyms of C.fimbriata (Upadhyay 1981, Webster and Butler 1967,Zimmermann 1900).

This study applies analyses of allozymes, DNAsequences, interfertility tests, host specialization andmorphology to identify cryptic species among isolates

of the C. fimbriata complex from North Americausing the phylogenetic species concept (Harringtonand Rizzo 1999). This concept recognizes species aspopulations or lineages with unique phenotypiccharacters, such as morphology and host specializa-tion.

METHODS AND MATERIALS

Fungal isolates.—A limited number of isolates fromPopulus tremuloides (aspen), Prunus dulcis (almond),Carya spp. (hickory) and Quercus spp. (oak) wereobtained from culture collections and plant patholo-gists. The second author collected isolates from C.cordiformis (bitternut hickory) in northeastern Iowa andP. dulcis in the Central Valley of California.

Further attempts were made during summer 2001 tocollect isolates of Ceratocystis spp. from Iowa. We examinedand obtained isolates from recently wounded Quercus spp.at the Yellow River State Forest in northeastern Iowa. Anadditional isolate was obtained from a bleeding canker onthe European species Q. robur in an experimental plantingat the Iowa State University (ISU) research farm at Rhodes.

Lineagea

Isolate orspecimennumberb

GenBankAccession No.

Additional isolate (specimen)numbers/collectorc Source

CollectionLocation

C1822 AY 907042 CBS 114717, BPI 843734/Harrington

P. dulcis, canker California

C1837 Johnson Quercus rubra, wound IowaC1841 Johnson Prunus serotina, wound IowaC1953d Johnson Populus grandidentata,

woundIowa

C1954 AY 907043 Johnson Tilia americana, wound IowaC1955d Johnson Quercus macrocarpa,

woundIowa

C1956d Johnson Q. macrocarpa, wound IowaC1957d Johnson Celtis occidentalis, wound IowaC1958d Johnson Carya ovata, wound IowaC1959d Johnson T. americana, bark beetle

(associated with wound)Iowa

C1963d Johnson Prunus serotina, wound IowaC1964 Johnson Q. macrocarpa, bark beetle

(associated with wound)Iowa

C2053 Harrington P. dulcis, canker CaliforniaCherry, Japan C1707 AY 907044 MCC - NIES 321/ Matsuya Betula platyphylla,

from logJapan

C1762 MCC- NIES 335/ Matsuya Betula platyphylla,from log

Japan

Cherry, WI C1965d AY 907045 Johnson Prunus serotina, wound Wisconsin

a Based on ITS rDNA sequences and/or allozyme analysis.b Isolate numbers preceded by C are from the collection of T. C. Harrington.c ATCC 5 American Type Culture Collection; CBS 5 Centraalbureau voor Schimmelcultures; MCC-NIES 5 Microbial

Culture Collection at National Institute for Environmental Studies, Heita, Kamaishi, Iwate, Japan; BPI 5 specimen from theU.S. National Fungal Collection.

d Extraction and analysis of allozyme electromorphs not replicated.

TABLE I. Continued

JOHNSON ET AL: CERATOCYSTIS FIMBRIATA: PHYLOGENY AND TAXONOMY 1069

TABLE II. Source of isolates from the Latin American and Asian clades of Ceratocystis fimbriata used in allozyme analysis

Isolatea

Additional numbers/collectorb Source Collection Location

C809 Capretti Platanus acerifolia ItalyC854 Clark Ipomoea batatas LouisianaC858 Harrington Platanus sp. CaliforniaC859 Harrington P. acerifolia CaliforniaC868 CMW2220 P. acerifolia FranceC918 Alfenas Gmelina arborea BrazilC925 CBS 115173, Alfenas G. arborea BrazilC940 CBS 152.62/Hansen Theobroma cacao Costa RicaC994 CBS 600.70/Figueiredo Mangifera indica BrazilC1004 CBS 153.62/Hansen T. cacao EcuadorC1022 Alvarez Citrus sinensis ColombiaC1024 Alvarez Coffea arabica ColombiaC1213 Somasekhara Punica granatum IndiaC1317 CBS 115162, Harrington Platanus occidentalis North CarolinaC1339 Britton P. occidentalis VirginiaC1345 Alfenas Eucalyptus sp. BrazilC1351 Harrington P. occidentalis KentuckyC1354 KFCF 9210/Kajitani I. batatas JapanC1355 KFCF 9001/Kajitani Ficus carica, from canker JapanC1391 IFO 30956/Kato F. carica JapanC1392 IFO32968/ Mukobata F. carica JapanC1393 IFO32969/ Mukobata F. carica JapanC1418 Cubeta I. batatas North CarolinaC1421 CBS 114723, Cubeta I. batatas North CarolinaC1442 CBS 115174, Alfenas Eucalyptus sp. BrazilC1451 Alfenas Eucalyptus sp. BrazilC1473 ICMP 894 I. batatas New ZealandC1475 ICMP 1731 I. batatas New ZealandC1476 ICMP 8579 I. batatas Papua New GuineaC1547 Paulin T. cacao Costa RicaC1548 CBS 114722, Paulin T. cacao Costa RicaC1551 Paulin C. arabica Costa RicaC1554 Alfenas M. indica BrazilC1587 Harrington T. cacao BrazilC1590 Harrington M. indica BrazilC1592 Harrington Annona sp. BrazilC1593 Harrington T. cacao BrazilC1597 Harrington T. cacao BrazilC1603 Harrington Manihot esculenta BrazilC1637 Harrington T. cacao Costa RicaC1641 Harrington Xanthosoma sp. Costa RicaC1642 Harrington Herrania sp. Costa RicaC1655 Baker M. indica BrazilC1657 Baker M. indica BrazilC1672 Baker Annona sp. BrazilC1673 Baker Eucalyptus sp. BrazilC1690 Harrington T. cacao EcuadorC1713 Harrington Hevea brasiliensis MexicoC1714 CBS 115164, Uchida Colocasia esculenta cv bunlong HawaiiC1715 CBS 114720, Uchida C. esculenta HawaiiC1717 CBS 114719, Uchida Syngonium sp. HawaiiC1774 Norman Syngonium sp. FloridaC1780 CBS 115165, Baker Xanthosoma sp. Costa RicaC1781 Harrington Syngonium sp. FloridaC1782 CBS 115166, Johnson Ficus carica Brazil

1070 MYCOLOGIA

At another site south of Boone, samples were taken froma C. ovata (shagbark hickory) tree with crown dieback anda C. cordiformis tree with recent attacks by a wood-boringbeetle (Agrilus sp.). Ceratocystis species were not isolatedfrom either hickory tree initially, but when the same treeswere resampled 2 wk later, the wounded bark was found tobe extensively colonized and isolates were recovered.Isolates also were obtained from C. cordiformis trees ina forest stand in Coggan, in northeastern Iowa, where anoutbreak of the hickory bark beetle (Scolytus quadrispinosusSay) was ongoing; isolates were obtained from beetlegalleries and from discolored wood associated with beetleattacks. Additional isolations were obtained from loggingwounds on C. cordiformis and P. serotina (black cherry) treesin the same stand. Another outbreak of S. quadrispinosuswas located near Cambria, in southern Iowa, and an isolatewas obtained June 2002 from a C. ovata tree infested withthe beetle.

In 2002 trees were wounded artificially at four locations inIowa. Wounds were created at breast height (1.4 m) on themain stem by removing a 6 3 6 cm patch of bark witha flame-sterilized hatchet, then bruising the bark with theback of the hatchet along two sides of the wound to loosenthe bark. Each site was revisited approximately 10 d afterwounding, and samples were taken from the wound faceand from bark surrounding the wound. Isolations weremade with the carrot disk method of Moller and DeVay(1968). At the first site, north of Ogden, one tree from eachof six species (P. serotina, Q. macrocarpa [bur oak], C. ovata,Celtis occidentalis [hackberry], Populus grandidentata [big-tooth aspen] and Tilia americana [basswood]) was woundedin mid-June. Isolates of a Ceratocystis sp. were recoveredfrom all but the Carya ovata and Celtis occidentalis trees. Ata second site near Lucas one tree each of five species waswounded at the end of June, and isolates of Ceratocystiswere recovered from Q. macrocarpa, Carya ovata and Celtisoccidentalis, but no Ceratocystis was recovered from Prunus

serotina and Ulmus rubra (red elm). The third site, nearAmes, contained both upland and bottomland species, andPrunus serotina, Q. macrocarpa, Carya ovata, Ostryavirginiana (ironwood), Juglans nigra (black walnut),Gleditsia triacanthos (honeylocust), Fraxinus pennsylvanica(green ash), and Populus deltoides (cottonwood) werewounded in mid-July, but Ceratocystis was isolated only fromO. virginiana. A fourth site north of Boone was visited inAugust, wounds were made on various hardwood species,and isolations were attempted 10 d later, but Ceratocystis wasnot recovered.

Representative isolates from the above collections wereselected for DNA sequence and allozyme analyses, in-oculation studies and mating experiments (TABLE I).Additional isolates representing the Latin American andAsian clades of C. fimbriata were used for allozyme analysis(TABLE II). Isolates C904, C1062 (CMW 4081), C1083(CMW 4110) and C1360 (JC 6885) of the African speciesC. albifundus were supplied by J. Roux and used as anoutgroup taxon in DNA sequence and allozyme analyses.

ITS sequencing and analysis.—Template DNA for PCRwas obtained from mycelium grown on 10 mL of broth(MYB, 2% malt extract and 0.2% yeast extract) atapproximately 24 C for 7–10 d, or from myceliumscraped from 1–2 wk old cultures grown on plates ofmalt yeast-extract agar (MYEA, 2% malt extract, 0.2%

yeast extract, 2% agar). DNA extraction was performedwith micropestles and microcentrifuge tubes followingthe method of DeScenzo and Harrington (1994). ThePCR primers, reagents and cycling conditions were aspreviously described (Harrington et al 2001). Sequenc-ing was performed at the ISU DNA Sequencing andSynthesis Facility using the PCR primers. Sequenceswere aligned manually by adding gaps, and parsimonyanalysis was performed with PAUP 4.0b10 (Swofford2002). Ceratocystis albifundus was used as outgrouptaxon, and the ingroup was considered to be mono-

Isolatea

Additional numbers/collectorb Source Collection Location

C1809 CBS 115167, Harrington Syngonium sp. FloridaC1810 Grillo Spathodea campanulata CubaC1811 Harrington S. campanulata CubaC1812 Harrington S. campanulata CubaC1817 CBS 114718, Harrington Xanthosoma sp. CubaC1848 Harrington F. carica BrazilC1849 Harrington F. carica BrazilC1859 Harrington Colocasia esculenta BrazilC1860 Harrington C. esculenta BrazilC1863 Harrington C. esculenta BrazilC1864 Harrington C. esculenta BrazilC1865 CBS 114713, Harrington C. esculenta Brazil

a Isolate number, those preceded by C are from the collection of Dr. Thomas C. Harrington.b ATCC: American Type Culture Collection; CBS 5 Centraalbureau voor Schimmelcultures; ICMP 5 Landcare Research New

Zealand ; CMW 5 Forestry and Biotechnology Institute, University of Pretoria, South Africa; IFO 5 Institute for Fermentation,Osaka, Japan; KFCF 5 from collection of Y. Kajitani, Fukuoka Agricultural Research Center, Fukuoka, Japan.

TABLE II. Continued


phyletic. Of 714 total aligned characters, including gaps,217 were ambiguously aligned and excluded from theanalysis, 110 remaining sites were variable, and of these,32 were parsimony informative. Except for gapsbetween ingroup and outgroup, there were onlysingle-base gaps, which were treated as a fifth character.A maximum parsimony heuristic search was performedwith all characters having equal weight. Starting treeswere obtained through stepwise addition, and tree-bisection-reconnection was used. Bootstrap analysis with2000 replications of heuristic searches was used todetermine support for internal branches.

Allozyme analysis.—One hundred ten isolates of C.fimbriata, including representatives from all three geo-graphic clades, and three isolates of the outgrouptaxon, C. albifundus, were tested for allozyme variation(TABLES I and II). Cultures were grown 14 d in 125 mLErlenmeyer flasks containing 30 mL of MYB at roomtemperature. Enzymes were extracted from mycelialmats onto paper wicks and stored at 280 C untilelectrophoresis (Zambino and Harrington 1992), whichwas performed on 12% starch gels (Harrington et al1996). Buffers and electrophoresis conditions areshown (TABLE III). With few exceptions (TABLE I),enzymes were extracted and tested for allozyme activityat least twice. Isolates C1418 and C1410 were includedin each gel as reference isolates. For each allozyme,electromorphs were designated by numbers in order ofdecreasing anodal migration, and the electromorphswere considered to be alleles. These data were used todevelop an uncorrected ‘‘p’’ distance matrix, andphenograms were generated with neighbor joiningand UPGMA (unweighted pair group method witharithmetic mean) with PAUP 4.0b10. The neighborjoining tree was rooted to C. albifundus.

Interfertility tests.—Ceratocystis fimbriata is both a hetero-thallic and a homothallic fungus, with two mating types;MAT-1 strains are self-sterile, but MAT-2 strains are self-fertile. The MAT-2 strains have both MAT-1 and MAT-2genes, but during unidirectional mating-type switching,the MAT-2 gene is deleted, and progeny that haveinherited nuclei with the deletion behave as MAT-1 andare self-sterile (Harrington and McNew 1997, Witthuhnet al 2000b). Thus MAT-2 (self-fertile) and MAT-1 (self-sterile) progeny are recovered from selfings of MAT-2strains. Most field isolates are MAT-2, and pairingexperiments are hampered in that MAT-2 strains areusually self-fertile. Some self-fertile MAT-2 strains pro-duce mutant sectors that lack the ability to produceprotoperithecia and perithecia, and these MAT-2sectors are self-sterile and function poorly as femalesin crosses (Engelbrecht and Harrington 2005, Harring-ton and McNew 1997, Harrington et al 2002).

MAT-2 tester strains that were used in pairings wereobtained by subculturing sterile sectors that arose sponta-neously from fertile, selfing isolates. The presence of theMAT-2 gene in these tester strains was confirmed with PCR(Witthuhn et al 2000b). All MAT-2 testers were self-sterile,except isolates C1959 and C1483, which produced de-formed perithecia that could be distinguished readily fromnormal perithecia produced in successful pairings withother strains. The MAT-1 testers were obtained by re-covering single ascospore progeny from self-fertile isolates.

The MAT-1 testers were used as recipients (females), andMAT-2 testers served as donors (males). MAT-2 testers alsowere paired with each other, but no combination resultedin an interfertile cross. Both male and female cultures weregrown on MYEA plates at room temperature. After 5 d maletesters were flooded with 15 mL sterile, deionized water andscraped with a sterile spatula to suspend spores and mycelialfragments. Female testers were 5 d old colonies, which

TABLE III. Abbreviations, buffer systems, and electromorphs found for 12 enzymes used in starch gel electrophoresis

Enzyme nameEnzyme

abbreviations EC numbersa Buffer systemsb

Electromorphsdiscerned

Acontinase ACN EC 4.2.1.3 MC 8.1 3Fumarate hydratase FUMH EC 4.2.1.2 MC 8.1 7Peptidase PEP EC 3.4.-.- MC 8.1 5Glucose-6-phosphate dehydrogenase G6PDH EC 1.1.1.49 MC 8.1 46-phosphogluconic dehydrogenase PGD EC 1.1.1.44 MC 8.1 5Aspartase aminotransferase AAT EC 2.6.1.1 MC 8.1 10NADH Diaphorase DIA EC 1.8.1.4 S-6 7Phosphoglucomutase PGM EC 5.4.2.2 S-6 3Mannose-6-phosphate isomerase MPI EC 5.3.1.8 S-6 5Adenylate kinase AK EC 2.7.4.3 S-11 8Malate dehydrogenase MDH EC 1.1.1.37 S-11 4Fluorescent esterase FE EC 3.1.1.- S-11 4

a Nomenclature Committee of the International Union of Biochemistry.b Buffer MC 8.1 was a continuous morpholine citrate system, adjusted to pH 8.1, run at 40 amps constant amperage for 4.5

hours. Buffer S-6 was a discontinuous Tris, citric acid system, adjusted to pH 8.6, run at 20 amps constant amperage for4.5 hours. Buffer S-11 was a discontinuous histine system, adjusted to pH 7.0, run at 40 amps constant amperage for 4.5 hours.Systems 6 and 11 modified from Soltis et al. (1983).

1072 MYCOLOGIA

received 1 mL of a conidial suspension from the male testerat the edge of the expanding female colony. Spermatizedcultures were allowed to grow 7 d at room temperature(24 C) before initial evaluation with a dissecting micro-scope. Ambiguous reactions were re-inspected after anadditional 4 d. When perithecia were found they wereexamined at 4003 or 10003 with a compound microscopeto see whether normal ascospores had formed.

Four host cross-inoculations.—We first tested whetherisolates from the four main, host-associated lineagesexhibited host specialization to four hosts: Populustremuloides, Q. rubra (red oak), Prunus serotina and Caryaillinoensis (pecan). Nine inoculation treatments, consist-ing of two isolates from each of the four major lineagesand a control, were applied to each host.

Lateral roots of P. tremuloides were dug from a clone nearJohnston, Iowa, and were maintained under mist until rootsprouts appeared (Benson and Schwalbach 1970). Sproutswere harvested when they were 3–6 cm tall, dipped in1000 ppm IBA (indolebutyric acid) and placed in peatpellets to root (Snow 1938). The plants were inoculated at 5–6 mo after rooting. Carya illinoensis seed (Sheffield’s SeedCompany, Locke, New York) were cold-stratified for 4 mothen germinated in Kimpak paper (Kimberly-Clark Corpora-tion, Irving, Texas) in a growth chamber set to a 16 h day(30 C)/8 hour night (20 C) cycle. After 14 d the germinatedseeds were transferred to 4-inch pots in the greenhouse andinoculated after 3–4 mo. Seed of Q. rubra (Sheffield’s SeedCo.) were cold-stratified for 6 wk, planted directly to 4-inchpots in the greenhouse, and the seedlings were inoculatedafter 2.5–3 mo. Half-sib seedlings of Prunus serotina were dugfrom under a tree near Ames, Iowa, planted in 4-inch pots,and grown 3–4 mo before inoculation.

Before inoculation, plants were grown on greenhousebenches in a mixture of 50% perlite, 50% Peat-lite mix(Fafard, Aawam, Massachusetts). All plants received a slowrelease fertilizer (Osmocote 19-6-12) at the time of sowingand biweekly feedings with liquid fertilizer (Miracle-GroEXCEL 21-5-20). Artificial light was used to maintaina 16 h day. Plants were transferred to growth chambers7 days before inoculation, where they were maintained ona 16/8 h light/dark cycle at 25 C. Each experiment (asingle host) was performed with a randomized completeblock design with six blocks, and six replications pertreatment. The C. illinoensis seedlings were inoculated ongreenhouse benches using the same experimental design.All experiments were repeated.

Inoculum was prepared from 7 d old cultures on MYEAplates (Baker et al 2003). Sterile water was added to theplates, the colonies were scraped and the suspension wasfiltered through four layers of sterile cheesecloth. Inoculumprimarily comprised endoconidia, which were adjusted to1.0 3 106 spores per mL with a hemacytometer. Controlinoculum was prepared by flooding sterile MYEA plates withwater, scraping and filtering the resulting solution throughsterile cheesecloth.

Plants were prepared for inoculation by making a downward-slanting horizontal cut through the bark and into the xylem ofeach stem with a sterile razor blade. Immediately after

wounding 0.2 mL of inoculum was introduced into the woundwith a syringe (21-gauge needle), and each wound was wrappedwith parafilm. Plants were watered daily, and any mortalityoccurring before the end of the experiment was recorded andthe plants harvested. Populus tremuloides and C. illinoensisplants were harvested at 40 d, while Q. rubra and Prunusserotina plants were harvested after 30 d. At harvest a shallowcut was made along the stem above and below the inoculationpoint, without cutting into the xylem, and the length ofcankers (phloem necrosis) was recorded. A slightly deeper cutthen was made, exposing the xylem tissue so that the totallength of xylem discoloration could be measured. The funguswas re-isolated from inoculated plants by placing discoloredtissue between carrot slices (Moller and DeVay 1968).

Length of xylem discoloration was analyzed by host plant,source of inoculum, experiment (within host), host 3

source interaction and source 3 experiment (within host)interaction using a multi-factorial analysis of variance(ANOVA) with controls excluded. For each inoculated hostANOVA indicated significant variation (P 5 0.05) due tothe two experiments (within host), so each experiment thenwas analyzed separately with one-way ANOVA. When theANOVA indicated significant variation among isolates(without controls), then Duncan’s multiple range test wasused to separate means, including the controls. Statisticswere performed with SAS statistical software (SAS Institute,Cary, North Carolina).

Prunus virginiana and Quercus macrocarpa cross-inocula-tions.—An additional experiment was performed inwhich two hosts, P. virginiana (common chokecherry)and Q. macrocarpa, were inoculated with isolates fromthe oak and cherry lineages: two isolates from the four-host inoculation experiment (C1009 from Quercus, andC821 from Prunus), two additional isolates from the oaklineage, six Iowa isolates from the cherry lineage, andisolate C1965 from P. serotina in Wisconsin.

One-year-old bareroot seedlings obtained from the IowaState Forest Nursery were grown in the greenhouse 6 wkafter bud break in 6-inch pots in greenhouse soil amendedwith Osmocote slow-release fertilizer. Inoculations wereperformed in the greenhouse as described above usinga completely random design with nine replications pertreatment. Plants were harvested 37 d after inoculation.Length of xylem discoloration for each inoculated host wasanalyzed separately as described above.

Host range of hickory isolates.—Four species of Carya andtwo species from the related genus Juglans wereinoculated with isolates from the hickory lineage.Bareroot seedlings (2 y old) of C. cordiformis, C. ovata,C. illinoensis, J. nigra (black walnut), and J. cineria(butternut) were grown in 2-gallon pots as describedabove and inoculated in a growth chamber with fourisolates of C. fimbriata collected from C. cordiformis.Inoculations were performed 67 d after planting (50 dafter first flush for Juglans spp. and 40 d after first flushfor Carya spp.). The experiment used a completelyrandomized design, with five replicates per treatment.Plants were harvested 6 wk after inoculation andevaluated for linear extent of xylem discoloration. The


experiment was repeated with the same hosts andprocedures, except that Carya plants were planted 7 dbefore the Juglans species in an effort to ensure thatplants would be in a similar growth stage at the time ofinoculation. Two-way ANOVA and Duncan’s multiplerange tests were used as described above.

Host range of aspen isolates.—Two inoculation experi-ments were performed in a growth chamber to test thesusceptibility of five Populus species to four isolates fromthe aspen lineage. In the first experiment, P. tremuloides,P. nigra (European black poplar), P. balsamifera (balsampoplar), and P. trichocarpa (black cottonwood) wereinoculated. The P. tremuloides plants were generatedfrom root sprouts and inoculated 3 mo later. All otherplants were 3 mo old rooted cuttings from dormanttwigs. The P. tremuloides sprouts were grown in 4-inchpots, while the other hosts were grown in 6-inch pots.Fertilization, care and inoculation procedures were asdescribed above. A completely randomized design wasused, with five replicates per treatment.

The experiment was repeated with an additional host, P.deltoides (eastern cottonwood). The P. tremuloides plants forthe second experiment were 9 mo old; the other hosts were4–5 mo old and were generated by rooting greenwoodcuttings. The first experiment was harvested at 7 wk, andthe second experiment was harvested after 5 wk. For eachexperiment, a two-way ANOVA was used to analyze theeffects of isolate, host and isolate 3 host interaction. Foreach experiment there was no significant interactionbetween host and isolate, so the results from the fourisolates were combined and one-way ANOVA was performedto compare the response of each host. Duncan’s multiplerange test was used to separate means in each experiment.

Morphology.—Isolates were grown on MYEA at roomtemperature (approximately 23 C) and lighting 5–12 dbefore measurements. Measurements of endoconidiaand endoconidiophores were made after 4–7 d growth,while perithecia and ascospores were measured after 7–10 d. Aleurioconidia were measured from cultures thathad grown 7–20 d. Material to be measured wasmounted in lactophenol cotton blue and observed withNomarsky interference microscopy (Olympus BH-2microscope), photographed with a Kodak DC 120digital camera and analyzed with Openlab digitalimaging software (Improvision Inc., Lexington, Massa-chusetts). Perithecia were measured with an eyepiecereticule at 2003 or 4003 magnifications. For moststructures 10 observations were recorded per isolate;when measuring endoconidia, however, 20 conidia weremeasured per isolate. Some structures were rare or hardto locate in a few isolates, and fewer observations weremade.

RESULTS

Phylogenetic analysis of ITS data.—Parsimony anal-ysis of aligned ITS rDNA sequences resulted in 138most parsimonious (MP) trees of 136 steps. The

other MP trees differed from that shown in FIG.1only in the minor branches, those withoutbootstrap support. Four host-associated lineageswere evident in all of the MP trees and ina neighbor joining analysis of the same dataset(not shown). All isolates from diseased Populusspp. grouped in a strongly supported branch.Similarly most isolates from Carya spp. groupedinto a single lineage with 99% bootstrap support(FIG. 1). Four Quercus isolates formed a separatelineage (89% bootstrap support) with a Betulaisolate from Japan and the holotype specimen ofC. variospora (BPI 595631) (FIG. 1). Nearly allPrunus isolates grouped into a single lineage (96%

bootstrap support) with a few isolates fromQuercus, Populus, Carya, Celtis occidentalis and T.americana, all wound associated; as well as a Betulaplatyphylla isolate from Japan. An isolate (C1965)collected from a wounded Prunus serotina tree inWisconsin did not group into any of the four host-associated lineages. These groups henceforth arereferred to as the aspen, hickory, oak, cherry andcherry-Wisconsin lineages.

Allozymes.—Forty-four electrophoretic phenotypes(ETs) were identified among the 113 isolates of C.fimbriata and C. albifundus tested. The neighborjoining phenogram, which was rooted to C.albifundus, and the unrooted UPGMA phenogramhad similar topologies (FIGS. 2 and 3). There wassubstantial variation in electromorphs amongisolates from the North American clade. TheLatin American clade was supported only weaklyin both analyses, and there was little variationamong the isolates tested. There was no bootstrapsupport for the Asian clade, but two distinctlineages were apparent, one comprising isolatesfrom Ficus carica in Japan and the other fromColocasia esculenta in the Pacific.

The four host-associated lineages within the NorthAmerican clade seen in phylogenetic analysis of ITS-rDNA data also were seen in the allozyme analyses.Isolate C1965 from Prunus serotina in Wisconsin wasunique (FIGS. 2 and 3). Six isolates of the hickory ITSlineage formed a sublineage within the hickorylineage, and these two sublineages are designatedhere as hickory A and hickory B.

Pairings.—Self-sterile MAT-2 tester strains fromthe cherry, hickory B, aspen and oak lineages wereused to spermatize MAT-1 testers of other repre-sentative isolates of the North American clade(TABLE IV). The MAT-2 testers formed successful,interfertile pairings (perithecia producing abun-dant, normal ascospores) only with MAT-1 testersfrom their respective lineages. When a MAT-2

1074 MYCOLOGIA

tester from the hickory B sublineage was used,perithecia and abundant ascospores were formedwith MAT-1 testers from both the hickory A andhickory B sublineages. MAT-2 testers from C578

and C856, which are from the cherry lineage,paired with most other cherry testers but did notpair with MAT-1 testers from T. americana.Conversely the MAT-2 tester from T. americana

FIG. 1. One of 138 most parsimonious trees based on ITS-rDNA sequences of Ceratocystis fimbriata isolates from the NorthAmerican clade. Consistency index (CI) 5 0.897, rescaled consistency index (RC) 5 0.863, retention index (RI) 5 0.962.Bootstrap values greater than 50% are indicated above the branches. The tree is rooted to C. albifundus.


mated with MAT-1 testers from the same host(C1954 and C1959) but not with other testers ofthe cherry lineage. The MAT-1 testers from theWisconsin Prunus isolate C1965 and the two

Japanese isolates failed to mate with any MAT-2tester.

Many pairings resulted in what appeared to behybrid perithecia with watery ascospore masses

FIG. 2. Neighbor joining tree of 113 isolates of C. fimbriata based on allozyme electromorphs of 44 electrophoreticphenotypes. The tree was rooted to C. albifundus. Bootstrap values greater than 50 are shown above branches.

1076 MYCOLOGIA

(TABLE IV). When observed microscopically theseperithecia were filled mainly with cellular debris,apparently from aborted asci and ascospores. Asco-spores, when present, were uncommon and often

misshapen. These pairings were interpreted as partialinterfertility due to interspecific crossing, consistentwith the interpretation of interspecific pairingsbetween other species of Ceratocystis (Engelbrecht

FIG. 3. Phenogram based on UPGMA analysis of 44 electrophoretic phenotypes from 113 isolates of C. fimbriata and threeisolates of C. albifundus. Bootstrap values greater than 50% are shown above branches.


and Harrington 2005; Harrington and McNew 1997,1998). Some pairings between testers also resulted insterile perithecia that lacked ascospores.

Four host cross-inoculations.—The analysis of vari-ance for the four-host inoculation experimentrevealed a significant effect on linear extent ofdiscoloration for each of the main factors, with thehost plant showing the greatest effect (F 5 35.92;P , 0.0001). Experiment (within host) was thesecond largest source of error (F 5 27.98; P ,

0.0001). There was also a significant host 3 source(of isolate) interaction (F 5 21.67; P , 0.0001).Consequently xylem discoloration then was ana-

lyzed separately for each host and each experi-ment.

Isolates of the aspen and hickory lineage causeddramatically more discoloration (FIG. 4) on P.tremuloides and C. illinoensis, respectively, than didisolates from the other lineages. Hickory isolate C682caused no more discoloration than the controls inboth Populus experiments and was suspected to havedeteriorated and lost pathogenicity. Thus hickoryisolate C684 was substituted for C682 in the in-oculation of other hosts.

Less evidence for host specialization was seen in theinoculations of Q. rubra and Prunus serotina. In thefirst Quercus inoculation, isolates from the oak and

TABLE IV. Pairings for sexual compatibility among tester strains from isolates of Ceratocystis fimbriata representing lineageswithin the North American clade

GenotypeMat-1,Female

MAT-2, Males

Cherry578(A)

Cherry856(B)

Cherry1959(A)

Hickory B1827(A)

Aspen995(A)

Oak1483(A)

Oak1843(A)

Aspen 1485 — Ha H H Ib — Sc

Hickory A 682 H H H I H H H684 H — H I — H H

1410 H H H I H H H1411 H H H I H He He

1828 H H H I H H H1844 H H H I H H H

Hickory B 1412 H H H I H H H1413 — — — I — — —1971 — — — I — — —

Oak 1009 — — — — — — I1483 — — — — — I I1843 — — — — — I I1846 — — — — — I I

Oak, Japan 1709 H H H H H H HCherry 578 I I H — — H —

821 I I H — — H —855 I I H — — H —856 I I H — — H —857 I I H — — H H

1822 I I H — — H —1841 I I — — — H S1953 I I H — — H H1955 I I H — — H —1957 — — — — — — —1964 I I H — — S S1954 H H I H H — —1959 H H I H H — —

Cherry, Japan 1707 H H H — — — —1762 H H H H H H H

Cherry, WI 1965 S — — S — — S

a H 5 hybrid: much cellular debris and few misshapen ascospores inside perithecium, exuded ascospore masses, whenpresent, watery in appearance.

b I 5 interfertile: ascospores abundant, with normal form; exuded ascospore masses white to peach colored.c S 5 sterile perithecia: perithecia produced, but no ascospores; perithecia often misshapen or poorly developed.

1078 MYCOLOGIA

cherry lineages and isolate C1412 from hickorycaused significantly more discoloration than did thecontrols and the other isolates (FIG. 4). When theexperiment was repeated, however, isolates of the oaklineage caused significantly more discoloration inQuercus than did the controls and the isolates fromthe other hosts, and the cherry isolates caused more

discoloration than did isolates from aspen or hickory.Similarly, when P. serotina seedlings were inoculated,isolates of the oak lineage caused discolorationsimilar to that caused by isolates from the cherrylineage in one experiment, but only the cherryisolates caused substantial discoloration in P. serotinain a second experiment. In both experiments on P.

FIG. 4. Average length of cankers (black bars) and xylem discoloration (open bars) caused by C. fimbriata isolates at 30–35 d after inoculation into Populus tremuloides, Carya illinoensis, Quercus rubra and Prunus serotina plants. Bars are means forsix replicates. Error bars represent standard error for xylem discoloration. Bars for xylem discoloration within a graph sharingthe same letter are not significantly different based on Duncan’s multiple-range test (P 5 0.05).


serotina isolates from the aspen and hickory lineagesinduced significantly less discoloration when com-pared to isolates from the cherry lineage.

Prunus virginiana and Quercus macrocarpa cross-inoculations.—Because of the ambiguous resultsseen when inoculating Q. rubra and P. serotina,another experiment was performed in which Q.macrocarpa and P. virginiana were inoculated withisolates from the oak and cherry lineages. For eachhost inoculations with all isolates resulted insignificantly greater discoloration than was seenin plants that received control inoculations(FIG. 5). Significant differences among the isolatesalso were seen when the controls were excludedfrom the analysis (Q. macrocarpa: F 5 2.51, P 5

0.0105; P. virginiana: F 5 3.27, P 5 0.0012).However there was no evidence of host specializa-tion, as isolates from each lineage producedsimilar amounts of discoloration in each host(FIG. 5).

Host range of hickory isolates.—When Carya spp. andJuglans spp. were inoculated with isolates from thehickory lineage, the ANOVA indicated that exper-iment, host and the host 3 experiment interactionall contributed significantly to the variation in

response; most of the variation was due to theinfluence of experiment (F 5 13.12, P 5 0.0004),and host was the next largest factor (F 5 6.95, P ,

0.0001). The isolates, two of the hickory Asublineage and two of the hickory B sublineage,were not a significant source of variation in thetwo experiments, nor was there any isolate 3 hostinteraction. Because of the great differences inresponse between the two experiments, they wereanalyzed separately with one-way ANOVA. In thefirst experiment (FIG. 6) the two Juglans specieshad significantly more discoloration than thethree Carya species. In the second experimentmore discoloration was seen in J. cineria than inany of the other hosts, but the amount ofdiscoloration was not significantly different thanthat seen in J. nigra, C. cordiformis, or C. illinoensis(FIG. 6).

Host range of aspen isolates.—Because the numberof hosts in the two experiments differed, eachexperiment was analyzed separately. Most of thevariation in the first experiment was explained byhost species (F 5 56.83, P , 0.0001), followed byisolate (F 5 15.42, P , 0.0001). However there wasno interaction between host and isolate (F 5 0.97,P 5 0.4755). The ANOVA for the second

FIG. 5. Average length of cankers (black bars) and xylemdiscoloration (open bars) in Quercus macrocarpa andPrunus serotina plants inoculated with C. fimbriata isolatesof the oak, cherry and cherry WI lineages. Errors barsrepresent standard error of xylem discoloration. Barssharing the same letter on a given host are not significantlydifferent based on Duncan’s multiple-range test (P 5 0.05)of xylem discoloration.

FIG. 6. Average length of xylem discoloration in threeCarya species and two species of Juglans inoculated withfour hickory-type isolates of C. fimbriata. Bars sharing thesame letter in a given experiment are not significantlydifferent based on Duncan’s multiple-range test. CC 5

Carya cordiformis; CO 5 Carya ovata; CI 5 Carya illinoensis;JN 5 Juglans nigra; JC 5 Juglans cineria.

1080 MYCOLOGIA

experiment was similar to that of the first, withhost species responsible for most of the variation(F 5 42.19, P , 0.0001). In experiment 2, isolatewas not a significant factor (F 5 0.25, P 5 0.8644)but there was a slight isolate 3 host interaction (F5 2.05, P , 0.0299). In the first experiment(FIG. 7) Populus trichocarpa was the most suscepti-ble host, followed by P. balsamifera; P. tremuloidesand P. nigra were less susceptible. The trend wassimilar in the second experiment, with significant-ly more discoloration in P. trichocarpa and P.balsamifera than in other hosts.

Morphology.—A wide range of perithecial sizes wasseen in the isolates measured. In general isolatesfrom sweet potato (C. fimbriata sensu stricto)produced smaller perithecial bases than isolatesfrom the North American clade, but perithecia ofisolates from Platanus spp. (C. platani) werecomparable in size to isolates of the NorthAmerican clade. Most perithecia of isolates fromthe North American clade produced a distinctcollar at the point where the perithecial neck

emerges from the base (FIG. 8), but this structurewas absent in isolates of C. fimbriata ss and C.platani.

Davidson (1944) and Hunt (1956) reported thatthe ostiolar hyphae of Ceratocystis variospora areshorter than those of C. fimbriata isolates, and wefound that the type specimen of C. variospora andisolates of the oak lineage had ostiolar hyphaeconsiderably shorter than those of isolates from theLatin American clade (TABLE V). Isolates of the cherrylineage and the cherry-WI isolate also had relativelyshort ostiolar hyphae, although only the cherry-WIisolate had ostiolar hyphae that were consistently asshort as those seen in isolates of the oak genotype.

Ascospores from North American isolates were 3.5–6.5 mm long and 3.0–5.0 mm wide (TABLE V), slightlysmaller than the range found in C. fimbriata ss and C.platani (5.5–7.0 3 3.5–5.5 mm). Hunt (1956) re-ported ascospores 4–6 3 2.5–3.5 mm in C. variosporaand 4.5–8 3 2.5–5.5 mm for C. fimbriata.

As is typical for Ceratocystis spp. the studied isolatesproduced two or three anamorphs, which areaccommodated in the genus Thielaviopsis (Paulin-Mahady et al 2002). Flask-shape phialides (and theendoconidia produced from them) of isolates in theNorth American clade were similar in dimension tothose reported by Hunt (1956) and Webster andButler (1967) for C. fimbriata and C. variospora.Isolates from the hickory A sublineage conspicuouslylacked flask-shape phialides (TABLE V).

All isolates in the North American clade produceda second endoconidial stage with doliiform conidiafrom wide-mouth phialides, but within the LatinAmerican clade only isolates from Platanus producedthese structures (TABLE V). This second type ofendoconidiophore often was found clustered aroundthe bases of perithecia produced in culture and insamples of naturally colonized plant tissue. Wide-mouth phialides were generally shorter (12–65 mm)than flask-shape phialides, similar to earlier reports(Webster and Butler 1967). Doliiform conidia were4.5–19.5 mm long 3 3.5 to 9.5 mm wide and oftenwere found in long chains. Webster and Butler (1967)reported that doliiform conidia ‘‘are at first hyaline,becoming subhyaline to light brown with age’’;however we observed a change in the color ofdoliiform conidia only among isolates from the aspenlineage. Doliiform conidia from the aspen lineagefrequently expanded in size after emerging from theirphialides and developed into thick-walled, melanizedchlamydospores.

Aleurioconidia were 8.5–26 mm long 3 6.5–17.5 mm wide, were produced blastically and accumu-lated in chains. No obvious differences were seen inthe size of aleurioconidia among isolates of the

FIG. 7. Average length of cankers (black bars) and xylemdiscoloration (open bars) in five Populus species inoculatedin two experiments with four aspen-type isolates C. fimbriataat 40 d after inoculation. Bars sharing the same letter withinan experiment are not significantly different based onDuncan’s multiple-range test of xylem discoloration. PT 5

Populus tremuloides; PB 5 Populus balsamifera; PC 5

Populus trichocarpa; PD 5 Populus deltoides.


1082 MYCOLOGIA

various lineages, but isolates from the hickory Asublineage did not produce aleurioconidia.

These differences in morphology are incorporatedinto the emended descriptions of C. variospora andthe newly recognized taxa in the North Americanclade.

TAXONOMY

Ceratocystis variospora (Davids.) C. Moreau, Reveu deMycologie, Suppl. Colonial 17:22. 1952. FIGS. 8–16; Endoconidiophora variospora Davids., Mycologia 36:303.

1944.; Ophiostoma variosporum (Davids.) Arx, Antonie van

Leeuwenhoek 18:212. 1952.

EMENDED DESCRIPTION: Cultures on malt yeast-extract agar hyaline at first with a fluffy appearance,becoming brown, gray or olive-green after 2–4 days,undersurface of agar turning dark; radial growth18 mm at 5 d; odor sweet, often with banana scent.Hyphae hyaline to pale brown, often terminating asendoconidiophores. Perithecia (FIG. 8) with basessuperficial to partially immersed, bases black or rarelybrown, globose, 130–350(425) mm diam, unorna-mented or with undifferentiated hyphae attached;possessing a collar at the base of the neck 51–80 mmwide; necks black, slender, up to 830 mm long, 25–50 mm diam at base and 12- mm at the hyaline tip;ostiolar hyphae (FIGS. 9, 10) hyaline, 10–20 innumber, 1–2 mm wide (Hunt), 22–50 mm long,tapering to a blunt tip; asci not seen; ascospores(FIG. 11) with outer cell wall forming a brim, hat-shaped, 3.5–6.0 3 3.0–5.0 mm. Endoconidiophores oftwo types; one flask-shape, hyaline to light brown,septate with conidiophores 52–198 mm long, conidio-genous cell 37–66 mm long, width 4.5–7.0 mm at baseand 2.5–4.5 mm at the mouth; producing hyalineendoconidia (FIGS. 12, 13) 6.0–30.0 3 2.5–5.0 mm; theother endoconidiophores shorter, 32–90 mm long,not tapering, often flared at mouth, conidiogenouscell 16–38 mm long, width (3.0) 4.0–5.5 mm at baseand 4.5–7.5 mm at mouth; producing doliiformendoconidia (FIGS. 14, 15), hyaline 5.5–10.0 3 5.0–8.0 mm. Aleurioconidia (FIG. 16) produced blastically,singly or in chains, orange-brown to brown, ovoid orobpyriform, smooth, 9.0–6.5 3 7.5–14.0 mm.

SPECIMEN EXAMINED: HOLOTYPE: USA. WEST

VIRGINIA: Moorefield, from cambium side of Quer-cus prinus bark, May 1943, M.E. Fowler, BPI 595631.

CULTURES EXAMINED: USA. MINNESOTA:Ramsey County, from sapwood of Q. ellipsoidalisstumps cut 2–3 wk previously, 1955 or 1956, R.Campbell, isolate C1009 (5 CBS 773.73, ATCC12861). Ramsey County, from sapwood of Q. ellipsoi-dalis stumps cut 2–3 wk previously, 1955 or 1956, R.Campbell, from isolate C1483 (5 ATCC 12866).IOWA: Harper’s Ferry, from wound on Q. alba stem,Jul 2001, J.A. Johnson, isolate C1843 (5 CBS 114715,BPI 843737). IOWA: Rhodes, from bleeding cankeron Q. robur, Sep 2001, J.A. Johnson, isolate C1846 (5

CBS 114714, BPI 843738).Comments: This species is similar to Ceratocystis

fimbriata sensu stricto (the sweet potato pathogen) butdiffers in the production of doliiform conidia fromwide-mouthed phialides, and it differs from C.fimbriata, C. cacaofunesta and C. platani in its shorterostiolar hyphae and slightly smaller ascospores. Thepresence of a distinct collar at base of perithecialnecks distinguishes C. variospora from C. fimbriata, C.cacaofunesta, C. platani, C. albifundus, and C.polychroma. Cultures of the recently described C.pirilliformis from Australia were not available at thetime of study, but the description by Barnes et al(2003) includes the presence of a collar at the base ofthe perithecial necks, as in C. variospora. Dimensionsof ostiolar hyphae were not given for C. pirilliformis,but the ostiolar hyphae illustrated were up to 60 mmlong (Barnes et al 2003), longer than those observedin isolates of C. variospora. Although C. variosporaand C. pirilliformis are morphologically similar, theITS sequences of isolates of C. pirilliformis are distinctfrom those of C. variospora (Thorpe et al 2005).Ceratocystis variospora differs from C. albifundus andC. moniliformis in the production of aleurioconidiaand from C. moniliformis in the absence of ornamen-tation on the perithecial bases.

Ceratocystis variospora was described by Davidsonbased on fruiting structures found on the inner barkof chestnut oak (Quercus prinus) in West Virginia1 wk after the bark was removed from a living tree(Davidson 1944). It also has been collected from Q.ellipsodalis stumps in Minnesota (Campbell 1957),from a wound on Q. alba in Iowa and from a bleedingcanker on Q. robur, also in Iowa. Isolates from

r

FIGS. 8–16. Ceratocystis variospora. 8. Perithecium. 9, 10. Ostiolar hyphae and emerging ascospores. 11. Ascospores. 12.Flask-shape endoconidiophore producing cylindrical endoconidium. 13. Cylindrical endoconidia. 14. Wide-mouthendoconidiophore with emerging doliiform endoconidium. 15. Doliiform endoconidia in a chain. 16. Aleurioconidium. Allfeatures from isolate C1009 except FIG. 10, which was from isolate C1822. Bars: 8 5 100 mm; 9, 10,12, 14 5 20 mm; 11 5 5 mm;13, 15, 16 5 10 mm.


Minnesota and Iowa had similar ITS sequences andwere sexually interfertile, and the ITS sequenceamplified from DNA extracted from the holotypespecimen of C. variospora was also similar. Amorphologically similar isolate (C1709) from a sporu-lating mat on a log of Betula platyphylla in Japan hasa similar ITS sequence, but the Japanese isolate is notinterfertile with the oak isolates from North America.Other isolates from wounds of various hardwoods inIowa and a Prunus sp. in Wisconsin were morpholog-ically indistinguishable from the Quercus isolates of C.variospora, but they differed in ITS sequence andallozyme electromorphs and the Quercus isolates werein a separate intersterility group. Most of theseisolates from hosts other than Quercus formedostiolar hyphae longer than 50 mm, longer than thosefound in the Quercus lineage of C. variospora, but noostiolar hyphae longer than 50 mm were seen in thePrunus isolates C1822 and C 1841. Also host-specialization of isolates to Quercus spp. or Prunusspp. could not be demonstrated clearly. For thepresent, all these isolates are considered C. variospora,but the emended description of C. variospora is basedsolely on the Quercus isolates.

Ceratocystis populicola J. A. Johnson and Harrington,sp. nov. FIGS. 17–25Culturae glycosmae, saepe bananae similes. Peri-

thecia basibus atris, globosa, 110–275 mm diam,collari basim colli circumdante; collum usque ad665 mm longum, diametro ad basim 24–45 mm et adapicem 13–30 mm; hyphae ostioli hyalinae, 42–75 mmlongae. Ascosporae 4.5–6.5 3 3.0–5.0 mm. Endoconi-diophora hyalina ad fusca, formis duabus; formaprima cellulaconidiogena ampulliformi apicem versusangustata, endoconidiis cylindricis 10–33 3 2.0–5.0(5.5) mm; altera forma: cellula conidiogena bre-viore, saepeapicem versus dilatata. Endoconidiisdoliiformibus, primo hyalinis, 6.5–12.0 3 3.5–8.5 mm, saepe tumidescentibus et fuscescentibus,crassitunicatis, 8.0–13.5 3 6.0–10.5 mm. Aleurioconi-dia singula vel catenata, cinnamomea vel brunnea,ovoidea vel pyriformia, levia, 9.0–18.5 3 8.0–17.5 mm.

Cultures on malt yeast agar hyaline to whiteinitially, becoming darker, and turning brown orolive-green after 2–4 d, radial growth 17–21 mm at5 d; cultures smell sweet or of banana oil. Peritheciaon MYEA fully formed after 4–6 d, scattered onsurface of agar or with bases partially submerged.Perithecia (FIG. 17) with black bases, globose, 110–275 mm diam; unornamented or with undifferentiat-ed hyphae attached, possessing a collar at the base ofneck, necks black, emerging from collars, hyaline attip, slender, up to 665 mm long, 24–45 mm diam atT

AB

LE

V.

Dia

gno

stic

feat

ure

so

fC

erato

cyst

isfi

mbr

iata

,C

.pla

tan

i,an

dth

eta

xare

cogn

ized

wit

hin

the

No

rth

Am

eric

ancl

ade

of

the

Cer

ato

cyst

isfi

mbr

iata

com

ple

x

Tax

on

Ho

sts/

Inse

cts

Co

llar

atb

ase

of

per

ith

ecia

ln

eck

Asc

osp

ore

s(m

m)

Len

gth

of

ost

iola

rh

yph

ae(m

m)

Fla

sk-s

hap

edp

hia

lid

esan

dcy

lin

dri

cal

con

idia

Wid

e-m

ou

thed

ph

iali

des

and

do

lifo

rmco

nid

ia

Mel

aniz

edd

oli

form

con

idia

(ch

lam

ydo

spo

res)

Ale

uri

o-

con

idia

C.

fim

bria

tase

nsu

stri

cto

Ipom

oea

bata

tas

No

5.5–

7.0

33.

5–5.

053

–136

Pre

sen

tA

bse

nt

Ab

sen

tP

rese

nt

C.

pla

tan

iP

lata

nu

ssp

p.

No

5.5–

6.5

34.

0–5.

057

–77

Pre

sen

tP

rese

nt

Ab

sen

tP

rese

nt

C.

vari

ospor

ao

akli

nea

geQ

uer

cus

spp

.Ye

s3.

5–6.

03

3.0–

5.0

22–5

0P

rese

nt

Pre

sen

tA

bse

nt

Pre

sen

t

C.

vari

ospor

ach

erry

lin

eage

Var

iou

sh

ard

wo

od

sYe

s4.

0–6.

03

3.0–

5.0

21–8

7P

rese

nt

Pre

sen

tA

bse

nt

Pre

sen

t

C.

pop

uli

cola

Pop

ulu

ssp

p.

Yes

4.5–

6.5

33.

0–5.

042

–75

Pre

sen

tP

rese

nt

Pre

sen

tP

rese

nt

C.

cary

ae

Pri

mar

ily

Cary

asp

p.

Yes

4.0–

6.0

33.

5–4.

532

–79

Pre

sen

tP

rese

nt

Ab

sen

tP

rese

nt

C.

small

eyi

Cary

asp

p./

Sco

lytu

squ

adri

spin

osu

s

Yes

4.0–

6.0

33.

5–5.

054

–101

Ab

sen

tP

rese

nt

Ab

sen

tA

bse

nt

1084 MYCOLOGIA

base and 13–30 mm at hyaline tip; ostiolar hyphaehyaline, slender, tapered to a blunt tip, 42–75 mmlong (FIG. 18). Asci not seen; ascospores (FIG. 19) withouter cell wall forming a brim, hat-shape, 4.5–6.5 3

3.0–5.0 mm. Endoconidiophores of two types; oneflask-shaped, hyaline to light brown, septate withconidiophores 45–200 mm long, conidiogenous cell35–85 mm long, width 3.5–7.0 mm at base and 3.5–

4.5 mm at mouth; producing hyaline endoconidia 10–33 3 2.0–5.0 (5.5) mm (FIGS. 20, 21); the otherendoconidiophores shorter, not tapering, oftenflared at mouth; often produced in masses aroundperithecial bases (FIG. 17); conidiophores 17–95(125)mm long, conidiogenous cell 12–40 mm long; width3.5–6.0 mm at base and 3.5–8.5 mm at tip of con-idiogenous cell; producing doliiform endoconidia,

FIGS. 17–25. Ceratocystis populicola. 17. Perithecium. 18. Ostiolar hyphae. 19. Ascospores. 20. Flask-shaped endoconidio-phore and cylindrical endoconidia. 21. Cylindrical and doliiform endoconidia. 22. Wide-mouth endoconidiophore withemerging doliiform endoconidia. 23. melanized doliiform endoconidia, most mature conidium at right. 24. Melanizeddoliiform endoconidia attached to wide-mouth endoconidiophore. 25. Aleurioconidium. All features from isolate C685. Bars:17 5 100 mm; 18, 20, 22 5 20 mm; 19 5 5 mm; 21, 23, 24, 25 5 10 mm.


hyaline at first, 6.5–12.0 3 3.5–.5 mm (FIGS. 22), oftenbecoming swollen and melanized with thick walls(FIGS. 23, 24), 8.0–13.5 3 6.0–10.5 mm. Aleurioconi-dia (FIG. 25) produced blastically, singly or in chains,orange-brown to brown, ovoid or pyriform, smooth,9.0–18.5 3 8.0–17.5 mm.

HOLOTYPE: CANADA. QUEBEC: from Populustremuloides, E. Smalley, BPI 843723, from isolate C685(5 CBS 115161).

CULTURES EXAMINED: CANADA. QUEBEC:from Populus tremuloides, E. Smalley, isolate C685(5 CBS 115161). USA. SOUTH DAKOTA: BlackHills, from P. tremuloides, 1980, T.E. Hinds, isolateC89 (5 CO 301, 5 CBS 114725). COLORADO: fromP. tremuloides, T.E. Hinds, isolate C1485 (5 ATCC24096). POLAND. KORNIK: from canker on Populushybrid, Aug 1976, J. Gremmen, isolate C947 (5 ATCC36291). from canker on Populus hybrid, Aug 1976, J.Gremmen, isolate C995 (5 CBS 119.78).

Etymology. populicola, Latin 5 on Populus.Comments: This species is similar to C. fimbriata ss

but differs in the production of doliiform conidiafrom wide-mouth phialides. The distinct collar at thebase of perithecial neck distinguishes C. populicolafrom C. fimbriata, C. cacaofunesta, C. platani, C.polychroma and C. albifundus. C. populicola differsfrom C. variospora and C. pirilliformis in the pro-duction of chains of swollen, melanized chlamydos-pores from wide-mouth phialides. Ceratocystis populi-cola differs from C. albifundus and C. moniliformis inthe production of aleurioconidia and from C.moniliformis in the absence of ornamentation on theperithecial bases.

In our inoculations only isolates of C. populicolawere capable of causing disease in Populus spp.Distinctive, target-shape cankers caused by C. fim-briata have been noted on P. tremuloides in Minnesota(Manion and French 1967, Wood and French 1963),Pennsylvania, much of the western USA, includingAlaska (Hinds 1972, Hinds and Laurent 1978) andQuebec, Manitoba and Saskatchewan in Canada(Zalasky 1965). All these reports are believed to beof C. populicola. It is likely that the pathogen ispresent wherever Populus tremuloides naturally occurs.Hybrid poplars were found infected at plantations inPoland (Gremmen and de Kam 1977, Przybyl 1984b),and isolates from these plantations are C. populicola.An additional report from Quebec (Vujanovic 1999)describes C. fimbriata infecting rooted cuttings of P.balsamifera, a host found susceptible to C. populicolain our inoculations.

Ceratocystis caryae J.A. Johnson and Harrington, sp.nov. FIGS. 31–33

Culturae glycosmae, saepe bananae similes. Peri-thecia basibus atris, globosa, 135–340 mm diam,aliquando collari basim colli circumdante; collumatrum, gracile, usque ad 950 mm longum, diametro adbasim 25–52 mm et ad apicem 15–30 mm; hyphaeostioli hyalinae, gracile, 32–80 mm longae. Ascosporae4.0–6.0 3 3.5–4.5 mm. Endocondidiophora hyalina adfusca, formis duabus; forma prima cellulaconidiogenaampulliformi apicem versus angustata, endoconidiiscylindricis 8.5–27.0(43.0) 3 2.5–6.0 mm; altera forma:cellula conidiogena breviore, saepeapicem versusdilatata. Endoconidiis doliiformibus, hyalinis, 6.0–13.5(16.0) 3 5.5–9.5 mm. Aleurioconidia singula velcatenata, cinnamomea vel brunnea, ovoidea velpyriformia, levia, 9.0–21.5 3 8.5–16.5 mm.

Cultures on malt yeast agar hyaline to whiteinitially, becoming darker, and turning brown, grayor olive-green after 2–4 d, culture texture varyingfrom fluffy to felty, undersurface of agar turning dark.Cultures with a sweet scent, often smelling likebanana oil. Perithecia on MYEA fully formed after4–6 d; perithecia scattered or clumped on surface ofagar or with bases partially submerged. Peritheciawith bases black, globose or broadly obpyriform, 135–340 mm diam; unornamented or with undifferentiat-ed hyphae attached; occasionally with collar at apex48–103 mm wide; necks black, tapering to a hyalinetip, up to 950 mm long, 25–52 mm diam at base and15–30 mm at tip; ostiolar hyphae hyaline, slender,tapered to a blunt tip, 32–80 mm long. Asci not seen;ascospores with outer cell wall forming a brim, hat-shape, 4.0–6.0 3 3.5–4.5 mm. Endoconidiophores oftwo types; one flask-shaped, hyaline to light brown,septate with conidiophores 42–510 mm long, conidio-genous cell 33–80 mm long, width 3.8–7.5 mm at baseand 3.2–4.8 mm at the mouth; producing hyalineendoconidia 8.5–27.0(43.0) 3 2.5–6.0 mm (FIGS. 31,32); the other endoconidiophores shorter, nottapering, often flared at mouth; often produced inmasses around perithecial bases, conidiophores 40–100 mm long, conidiogenous cell 15–55 mm long;width 5.0–6.5(7.0) mm at base and 5.5–8.0 mm at tip ofconidiogenous cell; producing hyaline doliiformendoconidia, 6.0–13.5(16.0) 3 5.5–9.5 mm. Aleurio-conidia produced blastically, singly or in chains,orange-brown to brown, ovoid or pyriform, smooth,9.0–21.5 3 8.5–16.5 mm (FIG. 33).

HOLOTYPE. USA. IOWA: Coggan, from Caryacordiformis (Wangenh.) K. Koch, Aug 2001, J.A.Johnson, BPI 843735, from isolate C1829 (5 CBS114716).

CULTURES EXAMINED: USA. IOWA: Coggan,from Carya cordiformis, Aug 2001, J.A. Johnson, isolateC1829 (5 CBS 114716). Clayton County, from C.cordifomis, Sep 1998, T.C. Harrington, isolate C1412

1086 MYCOLOGIA

(5 BPI 843728). Clayton County, from C. cordifomis,Sep 1998, T.C. Harrington, isolate C1413. BooneCounty, from C. ovata, Jun 2001, J.A. Johnson, isolateC1827 (5 CBS 115168). Boone County, from C.cordiformis, Jul 2001, J. A. Johnson, isolate C1845.Ames, from Ostrya virginiana, Aug 2002, J.A. Johnson,isolate C1971.

Etymology. caryae, Latin 5 on Carya.Comments: This species is morphologically similar

to C. variospora but differs in the length of theostiolar hyphae. C. caryae differs from C. fimbriata ssin the production of doliiform conidia from wide-

mouthed phialides, and from C. fimbriata, C.cacaofunesta, C. platani, C. polychroma and C.albifundus in the presence of a collar subtendingthe perithecial neck. The doliiform conidia andaleurioconidia of C. caryae are larger than thosereported for C. pirilliformis (Barnes et al 2003). C.caryae differs from C. moniliformis in the absence ofornamentation on the perithecial bases. C. caryaelacks the melanized doliiform conidia seen in C.populicola. All isolates of C. caryae sensu stricto havebeen recovered from Carya spp., Ulmus spp. or Ostryavirginiana.

FIGS. 26–33. Ceratocystis smalleyi and C. caryae. 26. Perithecium. 27. Ostiolar hyphae. 28. Ascospores. 29. Wide-mouthendoconidiophore. 30. Doliiform endoconidia. 31. Flask-shape endoconidiophore. 32. Cylindrical endoconidia. 33.Aleurioconidia. 26–30 from isolate C684 from the holotype of C. smalleyi; 31–33 from C1829, the holotype of C. caryae.Bars: 26 5 100 mm; 27, 29, 30 5 20 mm; 28 5 5 mm; 31, 32, 33 5 10 mm.


Ceratocystis smalleyi J.A. Johnson and Harrington, sp.nov. FIGS. 26–30Culturae odore dulci bananae carentes. Perithecia

basibus atris, globosa, 100–300 mm diam, aliquandocollari basim colli circumdante; collum atrum, grac-ile, usque ad 570 mm longum, diametro ad basim 22–80 mm et ad apicem 15–40 mm; hyphae ostiolihyalinae, graciles, 55–100 mm longae. Ascosporae4.0–6.0 3 3.5–5.0 mm. Endocondidiophora hyalinaad fusca, uniformia, brevia, cellulaconidiogena saepedilatati versus apicem, endoconidiis doliiformibus,hyalinis, 7.5–13.5(16.0) 3 5.5–9.5 mm. Aleurioconidianon visa.

Cultures on malt yeast agar hyaline to whiteinitially, becoming darker and turning brown, grayor olive-green after 2–4 d, often with lighter coloredgray to white patches, undersurface of agar turningdark, many isolates sectoring readily. Radial growth21 mm at 5 d; cultures may have a sweet scent, but thebanana odor typical of C. caryae is absent. Peritheciaon MYEA fully formed after 4–6 d, often fruiting inconcentric rings; perithecia on surface or with basespartially submerged. Perithecia (FIG. 26) with basesblack, globose or broadly obpyriform, 100–300(350)mm diam; unornamented or with undifferentiatedhyphae attached; occasionally with collar at apex 42–73(85) mm wide; necks black, tapering to a hyaline tip,up to 570 mm long, 22–80 mm diam at base and 15–37 mm at tip; ostiolar hyphae (FIG. 27) hyaline,slender, tapered to blunt tip, 55–100 mm long. Ascinot seen; ascospores (FIG. 28) with outer cell wallforming a brim, hat-shape, 4.0–6.0 3 3.5–5.0 mm.Endoconidiophores (FIG. 29) of one type, nottapering, often flared at mouth; commonly producedin masses around perithecial bases, conidiophoresmulticellular, 35–105 mm long, conidiogenous cell22–65 mm long; width 4.0–6.0 mm at base and 4.0–7.5 mm at tip of conidiogenous cell; producingdoliiform to cylindrical hyaline endoconidia(FIG. 30), 7.5–31.5 3 4.0–7.5 mm.

HOLOTYPE. USA. WISCONSIN: Hickory Ridge,from Carya cordiformis, 1993, E. Smalley, BPI 843722,from isolate C684 (5 CBS 114724).

CULTURES EXAMINED: USA. WISCONSIN: Hick-ory Ridge, from Carya cordiformis, 1993, E. Smalley,isolate C684 (5 CBS 114724). La Crosse, from C.cordifomis, 1986, E. Smalley, isolate C682. Evansville,from C. ovata, 1993, E. Smalley, isolate C683. IOWA:Clayton County, from C. cordiformis, Sep 1998, T.C.Harrington, isolate C1410. Clayton County, from C.cordiformis, Sep 1998, T.C. Harrington, isolate C1411.Coggan, from C. cordiformis, Aug 2001, J.A. Johnson,isolate C1828. Coggan, from C. cordiformis, Aug 2001,J.A. Johnson, isolate C1839. Coggan, from C. cordi-formis, Aug 2001, J.A. Johnson, isolate C1840. Coggan,

from C. cordiformis, Aug 2001, J.A. Johnson, isolateC1842. Coggan, from C. cordiformis, Aug 2001, J.A.Johnson, from isolate C1844. Cambria, from C.cordiformis, Aug 2002, J. A. Johnson, isolate C1952.

Etymology. smalleyi, named after the late EugeneSmalley, who associated this fungus with Scolytusquadrispinosus and brought the new taxon to ourattention.

Comments: This species differs from C. caryae in theabsence of cylindrical conidia from flask-shapephialides and in the absence of aleurioconidia. Allisolates of C. caryae from wounded Carya spp. orOstrya virginiana are closely related to C. smalleyibased on ITS sequence analysis and allozyme bandingpatterns; they behave similarly in inoculation tests,and they appear to be sexually interfertile. Theisolates from wounds produce pink ascospore masses,while ascospore masses of C. smalleyi are white tocream. Perithecia of C. smalleyi do not consistentlyproduce a distinct collar at base of perithecial necks,but such swellings can be seen in at least someperithecia of all isolates, as they can in perithecia of C.caryae, C. variospora and C. populicola. EugeneSmalley first isolated the fungus from a tree thathad been attacked by the hickory bark beetle (Scolytusquadrispinosus), and he later made collections inassociation with the beetle from other locations inWisconsin (pers comm). We later collected isolatesfrom northeastern and south-central Iowa. Isolateshave been made from hickory bark beetle egggalleries, from stained wood surrounding galleriesand from discolored sapwood associated with beetleattacks from previous years.

DISCUSSION

Analyses of ITS-rDNA sequences and allozyme elec-tromorphs showed a great deal of variation in theNorth American clade of C. fimbriata and point to theexistence of four host-associated lineages. Althoughrelationships among lineages were not well resolved,there was general agreement between the ITS andallozyme analyses in delimiting the host-associatedlineages, as has been found in other studies ofCeratocystis species (Witthuhn et al 2000a). Pairingsbetween mutant MAT-2 tester strains that had lost theability to self and MAT-1 testers provided evidence ofmany biological species, but not all of these biologicalspecies are formally recognized in this study.

To delimit species under the phylogenetic speciesconcept supported by Harrington and Rizzo (1999),a lineage should have a unique combination ofphenotypic characters. The taxa in the NorthAmerican clade can be distinguished from the LatinAmerican clade of C. fimbriata by their slightly smaller

1088 MYCOLOGIA

ascospores and the collar present at the base of theperithecial necks. The taxa within the North Ameri-can clade are distinguished from each other bya number of minor morphological characters, pres-ence or absence of conidial states and by host range.Inoculation experiments distinguished some of thehost-associated lineages, with strong evidence for hostspecialization shown by isolates from the aspen andhickory lineages. Isolates from the oak and cherrylineages showed little to no evidence of hostspecialization, so these are retained as a singlespecies.

The name C. variospora is available for the oaklineage. Ceratocystis variospora originally was reportedin West Virginia on the inner bark of Q. palustriscollected for tanning and later was collected inMinnesota from a fresh stump of Q. ellipsoidalis(Davidson 1944, Campbell 1957). The ITS sequencegenerated from the holotype specimen was similar tothat of the Minnesota isolates and Iowa isolates fromQ. alba, a native tree, and Q. robur, a Europeanspecies. The isolate collected from Q. alba wasrecovered from a wound, while the isolate from Q.robur was isolated from a bleeding canker in a smallexperimental planting where many of the Q. roburtrees showed severe cankering.

We also are applying the name C. variospora to thelineage containing isolates from wounds on Prunusand other hardwood species. The cherry and oaklineages could be separated based on differences inITS sequences, allozymes and interfertility, but theycould not be consistently distinguished throughmorphology or host specialization. Isolates froma Tilia tree were typical of the cherry lineage in ITSsequence, allozymes and morphology but testers fromthese isolates were able to mate only with themselves.Isolates from Betula platyphylla logs in Japan also weremorphologically similar to USA isolates from the oakand cherry lineages of C. variospora, but they wereintersterile with the USA isolates and with each other.The ITS sequence and allozyme electromorphs of theWisconsin isolate from Prunus were unique, but thisisolate is morphologically indistinguishable from C.variospora and behaved similarly in inoculations ofQuercus and Prunus.

Previous observations of C. fimbriata in NorthAmerica have focused on mortality of infected trees,but C. variospora appears to be more common andoccur on more hosts as a relatively innocuous woundcolonizer. The cherry lineage of C. variospora appearsto be particularly common as a wound colonizer ona wide range of tree species in Iowa, although it may actas a tree-killing pathogen on almond and other exoticPrunus species in California (DeVay et al 1968, Molleret al 1969, Teviotdale and Harper 1991). Although no

member of the C. fimbriata complex had beenreported previously in Iowa, isolates of C. variosporawere readily collected from wounds, especially woundsmade early in the summer. New host records for the C.fimbriata complex include Carya cordiformis, C. ovata,Celtis occidentalis, Ostrya virginiana, Prunus serotina,Populus grandidentata, Quercus alba, Q. robur, Q. rubraand Q. macrocarpa. New host genera and familiesinclude Carya (Juglandaceae), Celtis (Ulmaceae) andOstrya (Betulaceae). The two Japanese isolates fromlogs of Betula platyphylla (Betulaceae) provided by H.Matsuya also represent a new host record for the C.fimbriata complex.

Isolates of C. populicola and C. caryae mated onlywith their respective testers, but the male tester of C.caryae mated successfully with MAT-1 females of bothC. caryae and C. smalleyi. We have been unable toobtain a MAT-2 tester of C. smalleyi to perform thereciprocal crosses. The ITS analysis failed to distin-guish C. caryae and C. smalleyi, although there wassome support from the allozyme analyses for separa-tion of these new taxa. We since have analyzedsequences from portions of the elongation factor-1aand b-tubulin-1 genes for the North American clade,and C. caryae and C. smalleyi appear as two well-resolved sister species in both of those gene trees(unpublished data).

Inoculations of Carya and Juglans spp. showed thatC. caryae and C. smalleyi have a range of potentialhosts within the Juglandaceae. Juglans nigra, J.cineria, and the Carya spp. are all susceptible to C.caryae and C. smalleyi, and only these two new speciesare pathogenic to Carya. It is interesting that thesetwo species behave similarly in inoculation studies,both specialized to members of the Juglandaceae, andthey appear to be fully interfertile, yet they differsubstantially in morphology. Ceratocystis caryae, witha single exception, was isolated only from Carya spp.that had not been infested by the hickory bark beetle,although some of the trees were within 5 m of beetle-infested trees with C. smalleyi. Since completion ofthis study, we have isolated C. caryae from a woundedUlmus sp.

Isolates of C. smalleyi were obtained from treesinfested by the hickory bark beetle, which is commonthroughout the eastern United States and has beenassociated with substantial hickory mortality, especial-ly in C. cordiformis (Felt 1914, St George 1929, Gangeand Kearby 1979). Ceratocystis smalleyi might playa significant role in this mortality. The association ofC. smalleyi with bark beetles is unique in the C.fimbriata complex, and this might be a newly divergedspecies with unique adaptations, perhaps evolvingfrom the more typical wound-colonizing C. caryae.The absence of the endoconidial state with narrow


conidia and flask-shape phialides, the absence ofaleurioconidia and the absence of fruity volatiles inculture are likely derived characters that somehow aidthe unique association of C. smalleyi with Scolytusquadrispinosus. The three other species of Ceratocystisassociated with bark beetles also show a loss ofproduction of the narrow endoconidial state anda loss of aromatic volatiles when compared to theirmore typical relatives (Harrington and Wingfield1998).

Although only a limited number of isolates werestudied we conclude that Ceratocystis canker onaspen is caused by C. populicola. Published reportsshow that Ceratocystis canker on aspen occurs overa broad range, including neighboring Minnesota(Manion and French 1967, Hinds 1972, Hinds andLaurent 1978). However no cankers typical of thosecaused by C. populicola on aspen were observed inIowa and no isolates of C. populicola were recoveredduring two summers of collecting. Our inoculationswith isolates of C. populicola found that Populustremuloides, the common host of C. populicola, hada less dramatic response to inoculation than did P.balsamifera, which has been reported as a host onlyonce (Vujanovic 1999), or P. trichocarpa, which hasnot been reported as a host in North America.Populus deltoides and the European P. nigra also wereshown to be susceptible to C. populicola in ourinoculations. Given the susceptibility of many poplarspecies, the wide range of P. tremuloides, and the factthat C. fimbriata has been reported from much of thisrange (Manion and French 1967, Hinds 1972, Hindsand Laurent 1978), it is surprising that only P.tremuloides and P. balsamifera have been reported asNorth American hosts.

An outbreak of Ceratocystis canker on experimen-tal plantings of hybrid poplars occurred in the late1970s and early 1980s in Poland (Gremmen and deKam 1977; Przybyl 1984a, b). Two isolates fromPopulus species in Poland were identified clearly asC. populicola based on ITS sequences, allozymes andmorphology. It is likely that C. populicola is in-digenous to North America and was introduced toPoland on infected poplar cuttings. Consistent withinoculation tests, Przybyl (1984a) found that clonesfrom P. nigra were less susceptible than clones of P.trichocarpa.

Intersterility barriers have arisen within the NorthAmerican clade of C. fimbriata, and some populationswithin the clade have begun to diverge geneticallyand phenotypically. Two host-associated lineages havebeen defined here as new species. Ceratocystis smalleyiassociated with the bark beetle Scolytus quadrispinosusappears to have lost some spore stages and aromaproduction, and it may be a newly diverged species,

still sexually compatible with C. caryae. Within C.variospora, the oak lineage, the cherry lineage, theTilia genotype and the cherry-Wisconsin isolate mayrepresent populations undergoing speciation andmay prove to be true species.

ACKNOWLEDGMENTS

This research was supported by the National ScienceFoundation through grants DEB-987065 and DEB-0128104. Eugene Smalley, H. Matsuya, J. Katijani, J. Uchidaand others kindly provided isolates. We thank R. Hall forproviding Populus clones and John Nason for use of hislaboratory, advice and assistance with allozyme analysis. Wealso thank D. McNew, J. Steimel, D. Thorpe and A. Johnsonfor technical assistance.

LITERATURE CITED

Baker CJ, Harrington TC, Krauss U, Alfenas AC. 2003.Genetic variability and host specialization in the LatinAmerican clade of Ceratocystis fimbriata. Phytopatholo-gy 93:1274–1284.

Barnes I, Gaur A, Burgess T, Roux J, Wingfield BD,Wingfield MJ. 2001. Microsatellite markers reflectsimilar relationships between isolates of the vascularwilt and canker pathogen Ceratocystis fimbriata. MolPlant Path 2:319–325.

———, Roux J, Wingfield BD, Dudzinski MJ, Old KM,Wingfield MJ. 2003. Ceratocystis pirilliformis, a newspecies from Eucalyptus nitens in Australia. Mycologia95:865–871.

Benson MK, Schwalbach DE. 1970. Techniques for rootingaspen root sprouts. Tree Plant Note 21:12–14.

CAB International. 2001. Ceratocystis fimbriata [original textprepared by CJ Baker and TC Harrington]. CropProtection Compendium. Wallingford, UK: CAB In-ternational.

Campbell RN. 1957. Studies on the biology of some wood-staining fungi. University of Minnesota. 69 p.

Davidson RW. 1944. Two American hardwood species ofEndoconidiophora described as new. Mycologia 36:300–306.

DeScenzo RA, Harrington TC. 1994. Use of (CAT)5 asa DNA fingerprinting probe for fungi. Phytopathology84:534–540.

DeVay JE, Lukezic FL, English H, Trujillo EE, Moller WJ.1968. Ceratocystis canker of deciduous fruit trees.Phytopathology 58:949–954.

Engelbrecht CJB, Harrington TC. 2005. Intersterility,morphology and taxonomy of Ceratocystis fimbriatafrom sweet potato, cacao and sycamore. Mycologia97:57–69.

———, ———, Steimel J, Capretti P. 2004. Geneticvariation in eastern North American and putativelyintroduced populations of Ceratocystis fimbriata f.platani. Mol Ecol 13:2995–3005.

Felt EP. 1914. Notes on forest insects. J Econ Ent 7:373–375.

1090 MYCOLOGIA

Gagne JA, Kearby WH. 1979. Patterns of host tree visitationby Scolytus quadrispinosus (Coleoptera: Scolytidae). JKansas Ent Soc 52:112–118.

Gremmen J, de Kam M. 1977. Ceratocystis fimbriata, a fungusassociated with poplar canker in Poland. Euro J ForestPath 7:44–47.

Halsted BD. 1890. Some fungous diseases of sweet potato.The black rot. N.J. Ag. Exp. Sta. Bull. 76:7–14.

Harrington TC. 2000. Host specialization and speciation inthe American wilt pathogen Ceratocystis fimbriata.Fitopath Brasil 25S:262–263.

———, McNew DL. 1997. Self-fertility and uni-directionalmating type switching in Ceratocystis coerulescens,a filamentous ascomycete. Cur Genet 32:52–59.

———, ———. 1998. Partial interfertility among theCeratocystis species on conifers. Fung Genet Biol25:44–53.

———, ———, Steimel J, Hofstra D, Farrell R. 2001.Phylogeny and taxonomy of the Ophiostoma piceaecomplex and the Dutch elm disease fungi. Mycologia93:111–136.

———, Pashenova NV, McNew DL, Steimel J, KonstantinovMY. 2002. Species delimitation and host specializationof Ceratocystis laricicola and C. polonica to larch andspruce. Plant Dis 86:418–422.

———, Rizzo DM. 1999. Defining species in the Fungi. In:Worrall J J, ed. Structure and Dynamics of FungalPopulations. Dordrecht, Netherlands: Kluwer Academ-ic Press. p 43–71.

———, Steimel J, Wingfield MJ, Kile GA. 1996. Isozymevariation and species delimitation in the Ceratocystiscoerulescens complex. Mycologia 88:104–113.

———, Wingfield MJ. 1998. The Ceratocystis species onconifers. Can J Bot 76:1446–1457.

Hinds TE. 1972. Ceratocystis canker of aspen. Phytopathol-ogy 62:213–220.

———, Laurent TH. 1978. Common aspen diseases foundin Alaska. Plant Dis Report 62:972–975.

Hunt J. 1956. Taxonomy of the genus Ceratocystis. Lloydia19:1–58.

Manion PD, French DW. 1967. Nectria galligena andCeratocystis fimbriata cankers of aspen in Minnesota.For Sci 13:23–28.

Marin M, Castro B, Gaitan A, Presig O, Wingfield BD,Wingfield MJ. 2003. Relationships of Ceratocystisfimbriata isolates from Colombian coffee-growingregions based on molecular data and pathogenicity. JPhytopath 151:1–11.

Moller WJ, DeVay JE. 1968. Carrot as a species-selectiveisolation medium for Ceratocystis fimbriata. Phytopa-thology 58:123–124.

———, ———, Backman PA. 1969. Effect of someecological factors on Ceratocystis canker in stone fruits.Phytopathology 59:938–942.

Moreau C. 1952. Coexistence des formes Thielaviopsis etGraphium chez une souche de Ceratocystis major (vanBeyma) nov. comb. Remarques sur les variations desCeratocystis. Rev Mycol Suppl Colonial 17:17–25.

Paulin-Mahady AE, Harrington TC, McNew D. 2002.Phylogenetic and taxonomic evaluation of Chalara,

Chalaropsis and Thielaviopsis anamorphs associatedwith Ceratocystis. Mycologia 94:62–72.

Przybyl K. 1984a. Disease of poplar caused by Ceratocystisfimbriata Ell. et Halst. I. Isolation of C. fimbriata,symptoms of the disease and evaluation of resistance ofpoplar clones resulting from artificial infection. ArborKorn 29:89–103.

———. 1984b. Disease of poplar caused by the fungusCeratocystis fimbriata Ell. et Halst. II. Morphology of thepathogen. Arbor Korn 29:105–117.

Roux J, van Wyk M, Hatting H, Wingfield MJ. 2004.Ceratocystis species infecting stem wounds on Eucalyp-tus grandis in South Africa. Plant Path 53:414–421.

———, Wingfield MJ, Bouillet JP, Wingfield BD, AlfenasAC. 2000. A serious new wilt disease of Eucalyptuscaused by Ceratocystis fimbriata in Central Africa. ForPath 30:175–184.

Snow Jr AG. 1938. Use of indolebutyric acid to stimulate therooting of dormant aspen cuttings. J For 36:582–587.

Soltis DE, Haufler CH, Gastony GJ, Darrow DC. 1983. Starchgel electrophoresis of ferns: a compilation of grindingbuffers, gel and electrode buffers and staining sched-ules. Am Fern J 73:9–27.

Steimel J, Engelbrecht CJB, Harrington TC. 2004. De-velopment and characterization of microsatellite mar-kers for the fungus Ceratocystis fimbriata. Mol Ecol Note4:215–218.

St George RA. 1929. Weather, a factor in outbreaks of thehickory bark beetle. J Econ Ent 22:573–580.

Swofford DL. 2002. PAUP*: Phylogenetic analysis usingparsimony (*and other methods). Version 4.0b10a.Sunderland, Massachusetts, Sinaur Associates.

Teviotdale BL, Harper DH. 1991. Infection of pruning andsmall bark wounds in almond by Ceratocystis fimbriata.Plant Dis 75:1026–1030.

Thorpe DJ, Harrington TC, Uchida JY. 2005. Pathogenicity,internal transcribed spacer-rDNA variation and humandispersal of Ceratocystis fimbriata on the family Araceae.Phytopathology 95:316–323.

Upadhyay HP. 1981. A monograph of Ceratocystis andCeratocystiopsis. Athens, Georgia: University of GeorgiaPress. 176 p.

Van Wyk M, Roux J, Barnes I, Wingfield BD, Liew ECY, AssaB, Summerell BA, Wingfield MJ. 2004. Ceratocystispolychroma sp. nov., a new species from Syzgiumaromaticum in Sulawesi. Stud Mycol 50:273–282.

Vujanovic V. 1999. First report of Ceratocystis fimbriatainfecting balsam poplar. Plant Dis 83:879.

Webster RK, Butler EE. 1967. A morphological andbiological concept of the species Ceratocystis fimbriata.Can J Bot 45:1457–1467.

Wingfield MJ, DeBeer C, Visser C, Wingfield BD. 1996. Anew Ceratocystis species defined using morphologicaland ribosomal DNA sequence comparisons. Sys ApplMicrobio 19:191–202.

Witthuhn CR, Harrington TC, Steimel JP, Wingfield BD,Wingfield MJ. 2000a. Comparison of isozymes, rDNAspacer regions and MAT-2 DNA sequences as phyloge-netic characters in the analysis of the Ceratocystiscoerulescens complex. Mycologia 92:447–452.


———, ———, Wingfield BD, Steimel JP, Wingfield MJ.2000b. Deletion of the MAT-2 mating-type gene duringuni-directional mating-type switching in Ceratocystis.Cur Genet 38:48–52.

Wood FA, French DW. 1963. Ceratocystis fimbriata, the causeof a stem canker of quaking aspen. For Sci 9:232–235.

Zalasky H. 1965. Process of Ceratocystis fimbriata infection inaspen. Can J Bot 43:1157–1162.

Zambino PJ, Harrington TC. 1992. Correspondence ofisozyme characterization with morphology in theasexual genus Leptographium and taxonomic implica-tions. Mycologia 84:12–25.

Zimmermann A. 1900. Uber den Krebs von Coffeaarabica, verursacht durch Rostrella coffeae gen. etsp. nov. Bull. Inst. Bot. Gardens Buitenzorg 4:19–22.

1092 MYCOLOGIA

Phylogeny and taxonomy of the North American clade of the ... · pirilliformis I. Barnes & M. J. Wingf. from Australia (Barnes et al 2003) and Africa (Roux et al 2004) and C. polychroma

Documents